By using this site, you agree to our Privacy Policy and our Terms of Use. Close

Forums - Microsoft Discussion - TRUE XB1 vs PS4 spec comparison



Around the Network
walsufnir said:
fatslob-:O said:
ethomaz said:

Adinnieken said:

I think though in my original comment I acknowledged this.  The realworld bandwidth that Microsoft has achieved is 70-80% of the maximum theorethical.  And they state the reason why no system can achieve the maximum theorethical, not even Sony.  I don't disagree that theoretical maximum's aren't necessarily a figure we should use, but you can't say the PS4 will achieve 176GB/s and then say the Xbox One won't achieve its theorethical max when neither can reach it.

You are right but PS4 have more change to archive close the peak bandwidth than Xbone... in fact the eSRAM + DDR3 system have more bottlenecks than a single GDDR5 pool.

Why are you so worried about the the ESRAM ? It's 32 god damn megabytes! It's so tiny that it's almost a non factor to the system's bandwidth performance. 


Ok so we should also get rid of caches, right? Really, don't talk if you don't know what you are talking about.

It's cache! What do you think there used for ? 

Since when did shaders need fast access to the majority of the tasks they do ? 



fallen said:
Zappykins said:
Interesting, would you mind pointing to your sources? It will help answer some questions.

And I see it's already brought out some misinformation.

Curious, why are you pulling out one core on the PS4 for audio? It can also do some with 4 CU's of it's GPU (I know) and does have a little audio chip.


Well, I did it both ways, hence -9% raw.

PS4 audio chip just does compress/decompress.

Doing audio on CU's is apparently pretty darn difficult (difficult to synchronize due to high GPU latency).

Regardless, if they are forced to use up some of their 4 "extra" CU 's on audio, I think that's a win for X1. You would have to adjust those GFLOP numbers for PS4 down.

From what I've heard SHAPE is equal to about 1 core of audio processing after Kinect reserves. This is a VERY ballpark figure. SHAPE could be even more of a beast, but much of it (beyond the 1 core) is reserved for Kinect audio processing (which IMO sucks, I hate when X1 is weaker for Kinect's sake).

Ah, yes, ok, I see how you did that.  Yes, the extra audio chip isn't that powerful in the PS4.

I'm a little bewildered as to why they would use the GPU like that.  I'm trying to keep an open mind about it.  But it seems a bit like using miata to move pianos.  Miatas are great little fun sporty cars.  Good mileage, reliability,  but would be mismatched for moving pianos.

Any idea what some of the 'mystery' chips in the Xbox One are for?



 

Really not sure I see any point of Consol over PC's since Kinect, Wii and other alternative ways to play have been abandoned. 

Top 50 'most fun' game list coming soon!

 

Tell me a funny joke!

Zappykins said:
fallen said:
Zappykins said:
Interesting, would you mind pointing to your sources? It will help answer some questions.

And I see it's already brought out some misinformation.

Curious, why are you pulling out one core on the PS4 for audio? It can also do some with 4 CU's of it's GPU (I know) and does have a little audio chip.


Well, I did it both ways, hence -9% raw.

PS4 audio chip just does compress/decompress.

Doing audio on CU's is apparently pretty darn difficult (difficult to synchronize due to high GPU latency).

Regardless, if they are forced to use up some of their 4 "extra" CU 's on audio, I think that's a win for X1. You would have to adjust those GFLOP numbers for PS4 down.

From what I've heard SHAPE is equal to about 1 core of audio processing after Kinect reserves. This is a VERY ballpark figure. SHAPE could be even more of a beast, but much of it (beyond the 1 core) is reserved for Kinect audio processing (which IMO sucks, I hate when X1 is weaker for Kinect's sake).

Ah, yes, ok, I see how you did that.  Yes, the extra audio chip isn't that powerful in the PS4.

I'm a little bewildered as to why they would use the GPU like that.  I'm trying to keep an open mind about it.  But it seems a bit like using miata to move pianos.  Miatas are great little fun sporty cars.  Good mileage, reliability,  but would be mismatched for moving pianos.

Any idea what some of the 'mystery' chips in the Xbox One are for?

the problem with this line of thinking is audio on CU isnt difficult at all, and is fully supported in the SDK, and doesnt even require exclusive use of the CU.



it can see in total darkness!
try that on your ordinary camera!

Kinect=power



Around the Network

I love the debate! What really needs to be discussed though is the affordability of the two consoles. If you buy an X1 first or a PS4 first you can still buy the other. Unless you are a rampant fanboy of on or the other. Chances are good that the PS4 will get cheaper. Heck it's already just 400 bucks! I'm getting an X1 first but what's to stop me from getting a PS4? Spec talk is fine but it always boils down to a pi$$ing match. Just game people!

fallen said:
Oops realized I made a mistake in the OP and forgot to include the triangle setup rates in the PS4 part of the table.

Both PS4 and XB1 GPU have two triangle setup engines. However PS4's GPU operates at 800 mhz while X1's operates at 853 mhz. This means X1 can setup 7% more geometry over PS4.

Updated OP to include.


WUT?!!??!!? WUT!!?!?!?!!?!?!?!? Just what the fuck do you think shaders and tessellation are for? Do you even understand the way RAM is used for triangle setups? Like, WUT!?#!?$$?!%?#?$%?%!?$%?!



Zappykins said:
fallen said:
Zappykins said:
Interesting, would you mind pointing to your sources? It will help answer some questions.

And I see it's already brought out some misinformation.

Curious, why are you pulling out one core on the PS4 for audio? It can also do some with 4 CU's of it's GPU (I know) and does have a little audio chip.


Well, I did it both ways, hence -9% raw.

PS4 audio chip just does compress/decompress.

Doing audio on CU's is apparently pretty darn difficult (difficult to synchronize due to high GPU latency).

Regardless, if they are forced to use up some of their 4 "extra" CU 's on audio, I think that's a win for X1. You would have to adjust those GFLOP numbers for PS4 down.

From what I've heard SHAPE is equal to about 1 core of audio processing after Kinect reserves. This is a VERY ballpark figure. SHAPE could be even more of a beast, but much of it (beyond the 1 core) is reserved for Kinect audio processing (which IMO sucks, I hate when X1 is weaker for Kinect's sake).

Ah, yes, ok, I see how you did that.  Yes, the extra audio chip isn't that powerful in the PS4.

I'm a little bewildered as to why they would use the GPU like that.  I'm trying to keep an open mind about it.  But it seems a bit like using miata to move pianos.  Miatas are great little fun sporty cars.  Good mileage, reliability,  but would be mismatched for moving pianos.

Any idea what some of the 'mystery' chips in the Xbox One are for?

Please, oh please, ignore him before you get dragged into fantasy land, audio processing does NOT take that much resource from the GPU due to the nature of parallel processing. You'd have to be on crack to think that CPU and GPU are equal on the same tasks to start with since they are good at different things.



dahuman said:
fallen said:
Oops realized I made a mistake in the OP and forgot to include the triangle setup rates in the PS4 part of the table.

Both PS4 and XB1 GPU have two triangle setup engines. However PS4's GPU operates at 800 mhz while X1's operates at 853 mhz. This means X1 can setup 7% more geometry over PS4.

Updated OP to include.


WUT?!!??!!? WUT!!?!?!?!!?!?!?!? Just what the fuck do you think shaders and tessellation are for? Do you even understand the way RAM is used for triangle setups? Like, WUT!?#!?$$?!%?#?$%?%!?$%?!

He's likely right about the xbone being able to push more geometry since AMD doesn't impliment a parallel solution to primitive generation compared to the nvidia polymorph engine and instead has dedicated units for doing that stuff but all of that is useless since every other part of the xbone except for the cpu is significantly weaker. 



fatslob-:O said:
dahuman said:
fallen said:
Oops realized I made a mistake in the OP and forgot to include the triangle setup rates in the PS4 part of the table.

Both PS4 and XB1 GPU have two triangle setup engines. However PS4's GPU operates at 800 mhz while X1's operates at 853 mhz. This means X1 can setup 7% more geometry over PS4.

Updated OP to include.


WUT?!!??!!? WUT!!?!?!?!!?!?!?!? Just what the fuck do you think shaders and tessellation are for? Do you even understand the way RAM is used for triangle setups? Like, WUT!?#!?$$?!%?#?$%?%!?$%?!

He's likely right about the xbone being able to push more geometry since AMD doesn't impliment a parallel solution to primitive generation compared to the nvidia polymorph engine and instead has dedicated units for doing that stuff but all of that is useless since every other part of the xbone except for the cpu is significantly weaker. 

Only that doesn't even matter as much as today's engines are focused on different techniques vs the tiny % of difference. Instead of calling out triangles one by one for everything, people will just mesh it out and use tessellation or shader techniques instead. RAM management is the most important thing in this equation, not fucking piling triangles together at such a small % difference while wasting resources since the consoles don't have massive RAM for games in reality due to their ridiculous OS restrictions.